RAGCache: Efficient Knowledge Caching for Retrieval-Augmented Generation

[gaocegege]

https://arxiv.org/pdf/2404.12457

Co-design RAG 和 LLM inference